Comparison of gravitational wave detector network sky localization approximations
Comparison of gravitational wave detector network sky localization approximations
复制标题
DOI：
10.1103/physrevd.89.042004
复制
发表时间：
2013-10
影响因子：
5
通讯作者：
K. Grover;S. Fairhurst;B. Farr;B. Farr;I. Mandel;C. Rodriguez;T. Sidery;A. Vecchio
中科院分区：
文献类型：
--
作者：
K. Grover;S. Fairhurst;B. Farr;B. Farr;I. Mandel;C. Rodriguez;T. Sidery;A. Vecchio
在致密双星合并过程中发射的引力波是引力波探测器网络的一个有前途的来源。从引力波数据中推断源在天空中的位置的准确性是引力波天文学的几个潜在科学目标的限制因素，包括多信使观测。各种方法已被用来估计所提出的网络定位源的能力。在这里，我们比较了两种技术，用于预测天空定位的不确定性-定时三角测量和Fisher信息矩阵近似-与贝叶斯推理的完整的，连贯的数据集。我们发现，单独的定时三角测量往往高估的不确定性，在天空定位的一组信号从非旋转的紧凑型物体的二进制范围高达20 M_\odot$的总质量的中位数因子为4 $，和高估增加的系统的质量。我们发现，平均预测可以带来更好的协议，包括相位一致性信息的定时三角测量技术。然而，即使经过纠正，这些技术也会产生与特定模拟信号的全面分析显着不同的结果。因此，虽然近似技术在提供网络定位能力的快速、大规模估计中可能是有用的，但是完全相干的贝叶斯分析给出了针对个体信号的更鲁棒的结果，特别是在存在检测器噪声的情况下。
Gravitational waves emitted during compact binary coalescences are a promising source for gravitational-wave detector networks. The accuracy with which the location of the source on the sky can be inferred from gravitational wave data is a limiting factor for several potential scientific goals of gravitational-wave astronomy, including multi-messenger observations. Various methods have been used to estimate the ability of a proposed network to localize sources. Here we compare two techniques for predicting the uncertainty of sky localization -- timing triangulation and the Fisher information matrix approximations -- with Bayesian inference on the full, coherent data set. We find that timing triangulation alone tends to over-estimate the uncertainty in sky localization by a median factor of $4$ for a set of signals from non-spinning compact object binaries ranging up to a total mass of $20 M_\odot$, and the over-estimation increases with the mass of the system. We find that average predictions can be brought to better agreement by the inclusion of phase consistency information in timing-triangulation techniques. However, even after corrections, these techniques can yield significantly different results to the full analysis on specific mock signals. Thus, while the approximate techniques may be useful in providing rapid, large scale estimates of network localization capability, the fully coherent Bayesian analysis gives more robust results for individual signals, particularly in the presence of detector noise.